Setting up
Today we are going to take a look at the starwars database that is included in the tidyverse package.
library(tidyverse) # open our beloved package
starwars
New data wrangling functions: count(), pivot_longer(), pivot_wider()
Before moving to more advanced data wrangling functions, let’s revise the commands in the past recitations.
starwarsEdit <- ____ %>%
____(____) %>% # first, let's remove the columns films, vehicles, and starships; don't forget the pipe %>%!
group_by(_______) %>%
summarise(___________)# then, let's calculate mean, min and max values of height, mass, and birth_year by species and gender
Now, what if I want to know how many observations I have per gender and species?
starwars %>%
________(______________) %>% # by gender and species
summarise(n=n()) # use the n() function inside summarise()
Just so you know – there is a faster way to do this using count()
starwars %>%
count(____________) # insert the two groups here
Now we can play a bit with the original data and transform in the opposite format to the original one. We can do this by using the function pivot_longer() and pivot_wider(). Let’s have a look and see the arguments these functions take.
Let’s first select some of the columns so it doesn’t get too cluttery. Let’s select: name, height, mass, birth_year
starwarsSelect <- _______________ %>% ________(_____________________)
What kind of format is the dataset? Wide or long?
It’s wide indeed, so then let’s use pivot_longer() to make it longer.
_______________________ <- starwarsSelect %>% # identify the dataframe
pivot_longer(cols = _______________, # indicate the columns to pivot into longer format: that is, all columns except name
names_to = __________, # name of the column with the levels of the group; it should be in quotes
values_to = __________) # name of the column with the actual values; it should be in quotes
If you wanted, you could also go back to the original format by using pivot_wider():
starwarsSelectTransformW <- starwarsSelectTransformL %>%
pivot_wider(names_from = features, # name of the column to get the names from and transform into columns
values_from = values) # name of the column to get the values from
Nice uh? :)
Some more practice
Now, let’s import the dataset from exam 1. As you may recall, the dataset you were given in exam 1 reported real judgment data for grammatical and ungrammatical sentences.
exam1.data <- read.csv("exam1.data.csv")
What is the format of the dataset?
Whatever that was, let’s change it!
___________ <- exam1.data %>%
pivot_????( )
Aaaaand let’s change it back again!
___________ <- exam1.data %>%
pivot_????( )
If you compared the dataset newly wrangled with the original one, you would see that they are exactly the same!